In this paper, we propose resource allocation optimization problems for indoor Optical Wireless Communications uplink. More precisely, a Binary NP-HARD optimization problem together with its relaxation are proposed. Cuckoo Search (CS) and Genetic Algorithms (GA) have been unlikely to find feasible solutions when dealing with the relaxation of the problem. For the binary model, Binary Cuckoo Search outperforms the classical GA achieving better solutions in less computational CPU time.